Parallel Experimentation and Competitive Interference on Online Advertising Platforms

This paper shows the existence of bias in the estimation of ad effectiveness created by concurrent online experimentation by competing advertisers and proposes a framework to account and correct for it.